[rrd-users] DERIVE vs COUNTER

Jean-Yves Avenard jyavenard at gmail.com
Fri Mar 21 16:46:26 CET 2008


On Sat, Mar 22, 2008 at 2:12 AM, Alex van den Bogaerdt
<alex at ergens.op.het.net> wrote:
>  That's no reset.  Don't think of it that way, it will just confuse you.

I see why you think of it that way ;-)

>  This should be detectable. A suggestion follows below.

actually, the uptime value of the modem doesn't reset when there's
loss of DSL sync, so I can not easily test the way you are suggesting.

>  And, perhaps more important, the rate computed cannot be more than
>  12mbps.  Use this knowledge, that's why rrdtool know about a minimum
>  and maximum rate.

Ok, I understood that the maximum value entered was the maximum value
of the counter: so here 2^32.
So if I entered as maximum value 12mbit/s (so around 90000000: 1.5MB/s
* 60) it should detect the reset easily...

>  Using DERIVE, together with a minimum rate of zero, is a trick to
>  ignore *any* counter value which is lower than the previous one.
>  So, the question you need to ask yourself is: do you want to take
>  a few shortcuts at the expense of loosing a little bit of information.
>  If yes: go for derive + minimum_rate==0
>  If no: go for counter, and detect resets

hum.. i wouldn't use much as I am recording the counter every
minute... not much can be transferred during that timeframe
considering the low speed of the link.

looks like using DERIVE will be good enough for me.

Another case this is happening is on my FreeBSD router.
To measure some specific traffic, I created a MIB for snmp-net
allowing to read some firewall (ipfw) rules.
each time I edit the firewall rule (which can just be counters), it
resets all the counters.
There are no easy way to tell that the counters got reset and it
wasn't an overflow: uptime doesn't change in any ways.

thank you guys for your help and advices.
very much appreciated.


More information about the rrd-users mailing list